By Gordon Rugg
My latest book was published on April 30th. Blind Spot is the story of an ambitious idea. I wanted to develop a method to spot where experts go wrong when tackling difficult problems. What happened next includes: A mysterious undeciphered mediaeval manuscript; inventing a radically different way to handle online search; new approaches to forensic statistics; a different way of looking at the search for life on other worlds, and much more. None of that was quite what I’d expected…
I was particularly interested in one type of expert error. A lot of mistakes are very easy to spot: an aircraft crashes or a bridge collapses, and there’s a body count. Those have been studied in depth by disaster researchers. Other types of mistake, though, are much harder to spot, such as a research field getting bogged down because everyone in it was making the same faulty assumption, or because everyone was failing to spot a key point. That’s what I was interested in: invisible errors, where something failed to happen.
It’s a tricky area. You can easily show that most people – including experts in a wide range of fields – are really bad at formal reasoning. However, when you start looking at real-world problems, as opposed to exercises in classroom logic, you discover that it can be very difficult indeed to assess what the logically “correct” answer to a problem should be.
Blind Spot is about what happened when my colleagues and I joined together four “virtual toolboxes” of methods that each dealt with a different part of the problem. I saw where two of those toolboxes would fit together; Jo Hyde saw another key connection, and Sue Gerrard spotted the last one. We called the final product the Verifier approach.
One toolbox is about methods for gathering information from people. A second is about ways of visualising information in a way that accurately represents it. A third is about choosing the right type of logic to assess the evidence. Those three are about finding errors, and maybe finding promising new answers. The fourth toolbox complements the first three; it’s about choosing the right method to get information across to people with the minimum of error in the process.
How do you test an approach like Verifier? It was designed to tackle problems where the leading experts in a field had ground to a halt and couldn’t see a way forward. There were plenty of problems matching that description.
Our first test case was my work on the Voynich Manuscript, sometimes described as the world’s most mysterious book. It’s a handwritten book of about 240 pages, written in a unique script; since its discovery in 1912, it had never been deciphered, despite the efforts of some of the world’s greatest codebreakers. Within weeks, I spotted something that previous researchers had missed: a simple way of producing text with the same complex regularities as in the original manuscript, using nothing more complex than paper, pen and penknife.
Other cases followed. Sue Gerrard’s critical reassessment of the way researchers thought about autism was published in the top peer-reviewed journal in that field. The Search Visualizer software that I developed with Ed de Quincey could lead to a sea-change in searching the internet or large documents – from aircraft maintenance manuals or medical databases to the works of Shakespeare. Using Search Visualizer, David Musgrave and I spotted a literary structure in the book of Genesis that had apparently been missed by centuries of previous researchers. Search Visualiser showed up striking differences in gender-related language among Shakespeare’s plays. We are currently collaborating with forensics experts, an astrobiologist and medical researchers, all interested in new ways of visualising knotty problems and their solutions.
That’s the central story in Blind Spot. There’s a lot more on top of that.
It’s for sale on Amazon:
If you’d like to read more about the background, there are good articles here:
(An article in Wired magazine, by my book co-author Joe D’Agnese, about my work on the Voynich Manuscript)
(An article in The Tablet about the Voynich Manuscript by Batya Ungar-Sargon)
My article in Scientific American about my Voynich Manuscript work is here:
If you’d like to try out the Search Visualizer, there’s a free version available online here:
In coming weeks, we’ll be posting articles on this site that unpack some of the topics in Blind Spot in more detail.